>You could open the INCOMING.SDF file with the low-level function (FOPEN) and walk through it line by line. Check the first character of the line and INSERT it into the correct table. It would look something like this:
>
[snip]
> Case Left(lcRecord,1) = 'A' && A means this record goes into table1
> Insert Into Table1 Values SubStr(lcRecord,2) && Ecludes the first char
> Case Left(lcRecord,1) = 'B' && B means this record goes into table2
> Insert Into Table2 Values SubStr(lcRecord,2)
> ... && Do this for each record type
> ...
> EndCase
>EndDo
>FClose(lnHandle)

This would surely be the way to go, if it could work, but it won't - Insert works on field-by-field basis, so all SubStr(lcRecord,2) would have to be chopped into substrings, one per field, with additional work on thir conversion to proper data types (dates, numbers etc). Could take some hours to write. Append ... SDF does this automatically, provided the table structure matches the coulmn widths in the SDF file.

I've run a little test on the compatibility of new numeric data types with the SDF format - well, it works for itself, i.e. I've managed to copy from it into a SDF file, and append back (even though I have Set Point to "," - an European comma, which has always been an obstacle in previous versions of Fox - nice feature, MSFT!)

>>
>>use table1
>>append from incoming.sdf type sdf for c1="A"
>>use table2
>>append from incoming.sdf type sdf for c1="B"

I've suggested this simply because it was simple :)